Magnetic nanocomposites for organic compatible

miniaturized antennas and inductors

P. Markondeya Raj+, Prathap Muthana, T. Danny Xiao*, Lixi Wan, Devarajan Balaraman, Isaac Robin

Abothu, Swapan Bhattacharya, Madhavan Swaminathan and Rao Tummala

Packaging Research Center

813 Ferst Drive NW

Georgia Institute of Technology

Atlanta GA 30332-0560

+ raj@ece.gatech.edu Phone: 404 894 2652 Fax: 404 894 3842

* Inframat Corporation, 74 Batterson Park Road, Farmington, Connecticut 06032

Abstract: Current wireless systems are limited by RF technologies

in their size, communication range, efficiency and cost. RF

circuits are difficult to miniaturize without compromising

performance. Antennas and inductors are major impediments for

system miniaturization because of the lack of magnetic materials

with suitable high frequency properties. Keeping antenna and

inductor requirements into consideration, two magnetic

nanocomposite systems – silica coated cobalt-BCB and Ni

ferrite-epoxy were investigated as candidate materials.

Nanocomposite thick film structures (125-225 microns) were

screen printed onto organic substrates. Parallel plate capacitors

and single coil coplanar inductors were fabricated on these films

to characterize the electrical and magnetic properties of these

materials at low and high frequencies. Electrical characterization

showed that the Co/SiO2 nanocomposite sample has a

permeability and a matching permittivity of ~10 at GHz

frequency range making it a good antenna candidate. Both

polymer matrix composites retain high permeability at 1-2 GHz.

I. INTRODUCTION

Recent developments in IC and system-level packaging

technologies have lead towards smaller RF transceivers with

embedded passive components. Amongst the RF components,

antenna is still considered a major barrier for system

miniaturization. The effectiveness of conversion of electrical

power to electromagnetic power becomes practical only when

the antenna size becomes larger than an appreciable fraction of

the wave length. Low profile antennas suffer from relatively

narrow bandwidth, substrate dielectric loss, mutual coupling

with their substrate and surface wave perturbation issues. If

the material surrounding the antenna has high-permittivity and

permeability, the wavelength ( ) in such material will be

shorter by a factor of 1/ . Therefore, magnetic

nanocomposites can potentially lead to an order of magnitude

reduction in size because of their combined effect of high

permittivity and permeability. These dielectrics can increase

the path of the normal current and hence the radiation

efficiency. By tuning the nanocomposite formulation, the

material can be designed to have equal permittivity and

permeability in order to have good impedance match with that

of free space. The bandwidth of antennas on magnetic

substrate are also known to be wider than those on

conventional dielectric substrate. In this work, novel magnetic

nanocomposites were synthesized and investigated for their

magnetic permeability, dielectric constant and dielectric loss.

Reducing the particle size and the separation between

neighboring particles down to the nanoscale leads to novel

magnetic coupling phenomena resulting in higher permeability

and lower magnetic anisotropy. Co- or Fe-based

nanocomposites show much higher permeability than those

obtainable from the bulk Co or Fe metal. This large

enhancement in permeability is due to the interparticle

exchange coupling effect. The exchange interaction, which

leads to magnetic ordering within a grain, extends out to

neighboring environments (through either a spin polarization

or super-exchange interaction mechanism) within a

characteristic distance, lex. For Fe or Co, lex is 30 nm. The

exchange interaction in nanocomposites also leads to a

cancellation of magnetic anisotropy of individual particles and

the demagnetizing effect, leading to ultrasuperior soft

magnetic properties. By choosing a system with high

tunneling excitation energy, a huge increase in the resistivity

(from 10–6 Ohm-cm in pure metal to 109 Ohm-cm in

nanocomposite) can be achieved through the nanocompositing

technique. Because of the nanosized metal particles, the eddy

currents produced within the particle are also negligibly small,

leading to much lower loss for nanocomposites compared to

that of conventional microsized ferrites and powder materials.

Similarly, nanosized Ni ferrite particles are also expected to

have superior properties compared to their microsized counter

parts.

It is well known that most high permeability and high

permittivity materials have frequency-dependent properties.

This makes it difficult to design using these otherwise high-

performance materials. Hence, there is a strong interest to

0-7803-9085-7/05/$20.00 ©2005 IEEE

characterize the high-frequency properties of these materials.

Our previous work reported high frequency properties of thin

barium titanate, strontium titanate films and carbon black –

epoxy nanocomposites using multiline calibration method

with Coplanar Wave Guide structures [1]. Previous studies by

Zhang et al. [2] reported that while bulk Ni ferrite has unstable

magnetic permeability beyond 100 MHz, the Co-silica system

shows stable permeability till 240 MHz. In this work, Ni

ferrite-epoxy and Co-Silica/BCB systems were investigated

with parallel plate capacitor and inductor structures, for their

properties till 2 GHz. Based on these measurements, the

suitability of these materials for high frequency antenna and

inductor applications will be assessed.

II. SYNTHESIS AND FABRICATION

A. Silica-coated Co Synthesis: Inframat’s technology for the

synthesis and processing of magnetic nanocomposite is

different from the conventional metallic and ferrite materials.

The Co/SiO2 nanocomposie was synthesized using an aqueous

solution reaction technique, which is capable of synthesizing a

variety of nanostructured materials including metal, oxides,

and composites in large quantities. In the synthesis of

Co/SiO2 magnetic nanocomposite powders, experimental

procedures include preparing the starting precursors

containing Co and SiO2, co-atomization of the precursors to

form colloidal solution of Co and SiO2 and annealing the

colloidal solution to form Co/SiO2 nanocomposite powders.

A typical TEM bright field image for Inframat’s synthetic

Co/SiO2 nanocomposite is shown in Fig. 1. The TEM studies

revealed that the synthetic nanocomposite is a two-phase

material, where nanoparticles of Co are coated with a thin film

of silica. The Co phase has an average particle size of ~30

nm. Selected area electron diffraction experiments indicated

that the Co particles are fcc nanocrystals, where the matrix

silica phase is amorphous.

In order to study the microstructure in detail at the

nanometer level, localized regions at the Co/silica interface

have been studied using a microbeam diffraction technique.

The diffraction beam was reduced to approximately 10 nm in

size and diffracted at the area of interest. Two phases were

found in localized regions, including fcc Co and -phase SiO2.

Majority SiO2 coating is amorphous, as indicated in the XRD

and other selected large area diffraction techniques.

The Co-Silica nanoparticles (50 wt.% each) were mixed

with Benzo Cyclo Butene (BCB) polymer (Dow Chemicals,

Midland, Michigan) using conventional mixing. The as-

received BCB 3022-46 has 46 % volatile mesitylene. A paste

(60 wt. % inorganic and 40 wt. % as received polymer) with

suitable viscosity was then screen printed to a thickness of 225

microns onto high-temperature copper-clad FR-4 substrates

that are compatible with BCB curing at 250 C. Copper is

sputtered to a thickness of 3 microns, and patterned by an

etch-back process to form the top electrode.

B. Ferrite synthesis: In the ferrite nanoparticles synthesis,

again, a low temperature synthesis approach based on the

aqueous synthesis method has been developed to synthesis

very fine (Ni0.5Zn0.5)Fe2O4 nanoparticles. The procedures

include (1) preparation of a salt solution that contains Ni, Zn,

and Fe with the selected atomic ratio, (2) addition of the

NH4OH solution into the Ni and Fe precursor solution to

adjust pH, without any precipitation, (3) conversion of the

precursor solution into a Ni-Fe-O complex powder, and (4),

conversion (calcine) of the Ni-Zn-Fe-O material into

nanostructured (Ni0.5Zn0.5)Fe2O4 at low temperature in

hydrogen and oxygen controlled atmosphere.

In the microstructure studies, HRTEM specimens were

prepared by dispersing the powders in methanol. Drops of this

solution were then deposited on a carbon-grit and observed in

the microscope. Bright field images, electron diffraction, and

lattice images were carried out. A typical HRTEM atomic

resolution image shown the morphologies of the ferrite

nanoparticle materials is illustrated in Fig. 1. TEM studies at

atomic resolution and electron diffraction reveal that the

synthetic (NiZn)Fe2O4 nanoparticle is hexagonal. The size of

the ferrite nanoparticle is ~ 5 to 15 nm.

The nanoparticles of ferrite were dispersed into a

commercial epoxy to form a thick paste. The paste is then

screen-printed onto a FR4 substrate to a thickness of 125

microns. The composition of the paste consists of 50-70%

nanoparticles uniformly dispersed in an epoxy resin. The

epoxy can be cured to form a dense solid at temperatures

between 80-140oC. The curing time correspondingly varies

from 5 hrs to 30 minutes. In this study, the films were cured at

140 C for 30 minutes. The top electrode was sputtered to a

thickness of 3 microns and patterned by an etch-back process.

Fig 1. A typical TEM micrograph shows a two phase material,

where amophous SiO2 films are coated on the surface of Co

nanocrystals. A typical atomic resolution TEM micrograph

showing the microstructure of the (NiZn)Fe2O4 nanoparticles

Arrow points to the boundary between two nanoparticles.

III. ELECTRICAL CHARACTERIZATION

Parallel plate capacitor and single coil inductor structures

were fabricated to characterize the electrical properties of

these materials. The structure details are described in Fig. 2.

Different sized components were used to check the

consistency in the measurements and also for internal

calibration. To characterize the permittivity, both high

frequency and low frequency characterization techniques were

employed. Low frequency characteristics were measured with

a Precision LCR meter (Agilent 4285A). The dielectric

constants of both the nanocomposites are much higher than the

base polymer employed. The observed decrease in dielectric

0-7803-9085-7/05/$20.00 ©2005 IEEE

constant with frequency (Fig. 3) in metal-insulator composites

is indicative of the interfacial polarization relaxation.

Fig. 2. Fabricated inductor and capacitor structures. The outer

diameter for the inductor coil is 1.22 mm while the inner

diameter is 0.53 mm. The size of the capacitor is 2.33x 2.33

mm. For the capacitor, the adjacent metal is connected to the

bottom electrode.

The high frequency measurements were conducted with an

s-parameter Network Analyzer (HP Model 8720 ES). Two-

port measurements were employed to eliminate any systematic

errors that occur with single port measurements. The probes

were ground-signal type with a spacing of 500 microns.

Ground-Signal-Ground probes were not used in this study.

The transfer impedance (Z12 or Z21) can be estimated from the

s-parameters as:

21122211

12

11

250

SSSS

SZ (1)

for 2-port measurement with two probes located on the same

position of a coupon. Another form of this relationship is also

reported in the literature [3].

Fig. 3. Low frequency permittivity measurements of magnetic

nanocomposites.

A. Equations for parallel plate capacitor: Different size of

capacitors built on the same substrate, using same material and

processing can be used for internal calibration. Below the first

self-resonant frequency, a capacitor can be deemed equivalent

to a capacitor, inductor and resistor connected in series. The

equation for the serial circuit can be expressed as:

RCj

LjZ1

(2)

where Z is the impedance of the equivalent circuit. L, C, and R

are inductance, capacitance and resistance of the circuit

respectively. is the angular frequency. L, C, and R are

functions of frequency. Since there are three unknowns in the

equation different size capacitors would be needed. The goal

of this work is to extract permeability and permittivity. The

equations for the two capacitors will be:

1

11

1)Im(

CjLjZ (3)

2

22

1)Im(

CjLjZ (4)

The geometric part of the capacitance, Co, for parallel plate

capacitor can be estimated using as d

A0 where A is the

area of the plane and d is the distance of separation between

the planes. There is no simple equation for the inductance.

Generally, inductance with frequency dependent material

can be expressed as: 0)( LfL where L0 is the

inductance of an inductor filled with a non-magnetic

material. The L0 value can be estimated from the above

equations by plugging in the permittivity and permeability

measurements done independently at 50-100 MHz using

bulk inductors at Inframat Corporation [2]. Using the

estimated L0 and C0, the permeability and permittivity at all

the other frequencies can be solved from the simultaneous

equations (3-4) for two capacitors. Figs. 4 and 5 show the

permittivity and permeability values for both the

nanocomposite systems. It is interesting to note that both

materials retain a permittivity of more than 10 at GHz

frequency. The permeability value for Co-Silica-BCB is

close to 10 at 1 GHz frequency as seen in Fig. 5.

Fig. 4. High frequency permittivity of magnetic

nanocomposites.

B. Estimation of Ni Ferrite permeability from inductor

structures: Different size inductors were built on the

magnetic nanocomposite substrate as well as a glass

microslide with known permability of 1 and same thickness

as the magnetic nanocomposite (125 microns). For the

nanocomposite sample, the inductor is completely buried by

screen printing the paste above and below the structure, while

opening the pads for testing. Using same inductor structures

on glass and nanocomposite, the permeability was directly

estimated from the impedance measurements as:

0-7803-9085-7/05/$20.00 ©2005 IEEE

Glass

iteNanocompos

Z

Z

)Im(

)Im( (5)

It should be noted that the actual equivalent circuit of the test

structure is much more complicated, with an inductor and

resistor in series, and a capacitor in parallel. The resistive

effects, amplified by the loss, and capacitive effects,

amplified by the high permittivity nanocomposite,

significantly affect this estimate. Hence, this analysis is not

suitable for higher frequencies, where the Im(Z) Vs

frequency deviates from a straight line. The obtained

inductance as a function of frequency is plotted in Fig. 5,

along with that of Co-Silica-BCB for comparison. The

measurement indicates that the ceramic-polymer composite

retains a permeability of 2.5 at 1 GHz frequency.

Fig. 5. High frequency permeability of magnetic

nanocomposites.

High frequency characteristics of metal-based magnetic

nanocomposites were previously reported by Miura et al [4].

In their work, iron powder resin showed a permeability

decreasing from 3 to 2, as the frequency increases from 1

GHz to 5 GHz. Similarly, Ba2Ni2Fe12O22 showed a steep

decrease in permeability from 4.5 at 200 MHz to 2 at 1 GHz,

as studied by Shin and Oh [5]. While the permeabilities of

Fe72Al11O17 and Fe55Al18O7 are not stable beyond 1 GHz,

Co59Zr12O30 showed stable properties of 70-80 beyond 1

GHz. These are presumably not organic compatible

materials. Polymer based composites showed a steep

decrease in permeability from 10 at 1GHz to 1 at 3 GHz, as

reported by Yoshida et al [6]. The Co-Silica-Polymer system

showed higher permeability and permittivity than the

polymeric magnetic nanocomposite systems reported in the

literature because of its nanoparticle size and silica coating.

The high frequency stability of the Ni ferrite permeability is

always of concern, as is also seen in this study. The

nanocrystalline Ni ferrite is expected to yield higher

permeabilities than their bulk counterparts. One reason this

was not observed in these studies could be the use of lossy

polymer such as epoxy which could alter the inductive

behavior considerably. High performance BCB is currently

being used as the nanocomposite matrix to gain further

improvement in properties.

The dielectric loss of Co-silica-BCB varied with the filler

content. Incorporation of metal inside a polymer increases the

dielectric loss. In the nanocomposites studied here, the silica

coating is expected to prevent the charge leakage and help

lower the losses. The low frequency dielectric loss of these

nanocomposites varied from 0.003 for low filler content to

0.08 for higher filler content. Further improvements in the

formulation and dispersion methodologies are now being

investigated to lower the loss at high filler contents.

IV. CONCLUSIONS

Thick film Ni ferrite-epoxy and Co-silica-BCB

nanocomposites were characterized for their permittivity and

permeability over broad frequency range to assess their

suitability for antenna and inductor applications. Interestingly,

both materials show high permittivity and permeability. Co-

silica-BCB shows matching permittivity and permeability of

5-10 making it a good candidate for miniaturized antennas.

Fabricated inductor structures on Ni ferrite-epoxy showed that

the material retains a permeability of 2.5-3 at GHz frequency

range. Both material systems show significant advances over

existing organic (Printed Wiring Board) compatible magnetic

nanocomposites for high-frequency inductor applications.

ACKNOWLEDGMENT

This work was supported by the National Science

Foundation (NSF) through the NSF ERC in Electronic

Packaging (EEC-9402723) at Georgia Institute of Technology.

The authors also than Venky Sundaram (PRC, Georgia Tech.)

and Yong Kyu Yoon (ECE, Georgia Tech.) for their

assistance. Dr. T. Danny Xiao thanks Inframat Corp. for

funding his research activities in conducting this work.

REFERENCES

[1] D. Balarman, P. M. Raj et al., “ BaTiO3 films by low-

temperature hydrothermal techniques for next-generation

packaging applications”, Journal of Electroceramics, 13, 95-

100, 204.

[2] Y. D. Zhang, S. H. Wang, D. T. Xiao, J. I. Budnick and

W. A. Hines, “Nanocomposite Co/SiO2 soft magnetic

nanomaterials”, IEEE Trans. Magnetics, Vol. 37, no. 4, July

2001, pp. 2275-2277.

[3] Istvan Novak, Jason R Miller, “Frequency – Dependent

Characterization of Bulk and Ceramic Bypass Capacitors”,

Poster material for the 12th Topical Meeting on Electrical

Performance of Electronic Packaging, October 2003,

Princeton, NJ.

[4] T. Miura and S. Nakagawa, “Spectrum management of

pulse transmission by high-cut filter using magnetic loss”,

IEEE International Microwave Symposium, Phoenix, June

2001, Digest vol. 1, pp. 463-466.

[5] J. Y. Shin and J. H. Oh, “The microwave absorbing

phenomena of ferrite absorbers”, IEEE Trans. Magnetics, vol.

29, no. 6, Nov. 1993, pp. 3437-3439.

[6] S. Yoshida et al., “High-frequency noise suppression in

downsized circuits using magnetic granular films”, IEEE

Trans. Magnetics, vol. 37, no. 4, July 2001, pp. 2401-2403.

0-7803-9085-7/05/$20.00 ©2005 IEEE